**** BEGIN LOGGING AT Fri Sep 02 02:59:57 2011 Sep 02 04:08:55 hi. can anyone tell me what would cause a package to try to link to .../usr/lib/.libs rather than .../usr/lib/ ? Sep 02 04:14:58 anyone? Sep 02 06:07:28 03Stefan Herbrechtsmeier  07org.openembedded.dev * rcaf9b3cf27 10openembedded.git/recipes/gcc/ (gcc-4.3.3.inc gcc-4.5.inc gcc-package-cross.inc): (log message trimmed) Sep 02 06:07:28 gcc: Remove package libstdc++-dev and libgcc-dev from gcc-cross Sep 02 06:07:28 Remove package libstdc++-dev and libgcc-dev from gcc-cross as they Sep 02 06:07:28 already provided by gcc. Sep 02 06:07:28 Commit 8c42351519ece9ed60659a103e624d2e154b3641 add dev packages Sep 02 06:07:29 (libstdc++-dev libgcc-dev) to gcc-cross (gcc-package-cross.inc) Sep 02 06:07:30 which are already part of gcc. The libgcc-dev package both include Sep 02 06:12:02 I'm not sure if I understand. This happens when I am building the ace recipe. I get the error: "i486-angstrom-linux-g++: /home/.../build/tmp-angstrom_2008_1/sysroots/486sx-angstrom-linux/usr/lib/.libs/libstdc++.so: No such file or directory" Sep 02 06:12:37 I checked the 486sx-angstrom-linux/usr/lib directory, and the libstdc++.so is there, and I don't have a .libs directory Sep 02 06:31:56 darkopium: There are more people subscribed to the list. So I suggest to send a message to the list with all necessary information included. Sep 02 06:46:33 thanks, which list would that be? Sep 02 07:33:39 darkopium: Sep 02 07:34:15 darkopium: The lists are also listed on the Web site . Sep 02 09:02:47 pb_, ping? Sep 02 09:22:12 morning! Sep 02 09:25:03 morning Sep 02 09:25:27 XorA, did I fix what you asked about yesterday? Sep 02 09:28:43 ka6sox: I dont remembe reporting a bug :-) Sep 02 09:29:52 kk Sep 02 09:30:08 ka6sox: your working too hard if your fixing bugs I didnt report :-) Sep 02 09:30:09 I am searching logs now. Sep 02 09:30:28 ya, Sep 02 09:30:37 oh nm...I remember now Sep 02 09:30:40 6-14th Sep 02 09:31:17 minding our P's and Q's Sep 02 13:28:48 03Holger Hans Peter Freyther  07master * r639db8c766 10bitbake.git/lib/bb/fetch2/git.py: Sep 02 13:28:48 fetch2/git: Allow to specify the name of the checkout directory Sep 02 13:28:48 Signed-off-by: Holger Hans Peter Freyther Sep 02 13:28:48 Signed-off-by: Richard Purdie Sep 02 15:13:59 hi, I wonder how do_postinst is done within oe.dev, in oe-core it can be done at do_rootfs time or at runtime.... in oe.dev it's done only at runtime right? Sep 02 15:14:18 (it's for finding out the differences between oe.dev and oe-core) Sep 02 15:15:54 GNUtoo|work: btw even PATCHv3 has weird renamese.. Sep 02 15:15:56 renames Sep 02 15:16:55 ok Sep 02 15:16:59 what should I do then? Sep 02 15:17:16 I formated the patch that way: Sep 02 15:17:28 git format-patch HEAD -1 -M --find-copies-harder --subject-prefix="PATCH v3" Sep 02 15:18:05 should I send a v4? Sep 02 15:18:19 GNUtoo|work: there is extra file you don't want :) Sep 02 15:18:30 ah? let me look Sep 02 15:19:04 it was something like mtd-utils-1.4.6/mtd-utils-1.4.1/foo Sep 02 15:19:07 which file? Sep 02 15:19:31 sorry already deleted that email but noticed it inside Sep 02 15:19:37 ok Sep 02 15:19:43 let me look Sep 02 15:19:48 and thanks for the review Sep 02 15:20:36 yw Sep 02 15:23:08 hmmm Sep 02 15:23:10 diff --git a/meta/recipes-devtools/mtd/mtd-utils-1.4.1/add-exclusion-to-mkfs-jffs2-git-2.patch b/meta/recipes-devtools/mtd/mtd-utils-1.4.6/mtd-utils-1.4.1/add-exclusion-to-mkfs-jffs2-git-2.patch Sep 02 15:23:13 that's what it says Sep 02 15:23:20 but then it says it's 100% similar Sep 02 15:23:31 strange... Sep 02 15:24:09 I'd like to extend a bbclass in oe-core with tweaks in another layer Sep 02 15:24:22 What is the right way to do so? .bbappend ? BBCLASSEXTEND? Sep 02 15:24:53 JaMa|Wrk, btw I've also another patch: Sep 02 15:25:15 at eukrea we use ubifs, and had the exact same issue than the freerunner ubifs issue, at least it seems Sep 02 15:25:18 we fixed it Sep 02 15:26:19 so if the patch arrived, it should fix it for you too Sep 02 15:26:41 joelagnel, BBCLASSEXTEND seem for recipes Sep 02 15:28:00 I've seen it for classes in meta-oe: Sep 02 15:28:03 ../meta-openembedded/meta-oe/classes/gnome.bbclass:BBCLASSEXTEND += "native" Sep 02 15:28:35 GNUtoo|work, but it doesn't do what I want Sep 02 15:29:05 I can create a new class, inherit from the oe-core one and put my code there, but I think that'll get rejected ;) Sep 02 15:29:14 ok Sep 02 15:29:28 maybe ask in the list Sep 02 15:30:10 everyone here on long weekend holiday? ;) Sep 02 15:31:42 joelagnel: I don't think bbclassextend is what you want Sep 02 15:32:30 what tweaks do you want to apply? Sep 02 15:32:37 bluelightning, can I name a class in an upper layer with the same filename as that of a lower layer, and still be able to inherit the lower layer one? Sep 02 15:32:50 I don't think you can do it that way no Sep 02 15:33:03 if it's the same name it will just replace it effectively Sep 02 15:33:27 the most likely result in that case is you'll get a circular reference Sep 02 15:33:49 BlindMan, there was some talk about moving my changes (IMAGE_CMD_sdimg) to the bsp layer Sep 02 15:33:54 https://github.com/joelagnel/oe-core/blob/db4c960da57e8e0ae4741eea703a4e163c589efa/meta/classes/image_types.bbclass Sep 02 15:34:26 sorry I meant bluelightning ;) Sep 02 15:34:33 need coffee!!! :-D Sep 02 15:34:57 and need .bbclassappend ;) Sep 02 15:36:21 hmm; this is an interesting one Sep 02 15:36:49 I recall the discussion on the mailing list... I do wonder whether we ought to try to find a way to get this into oe-core itself Sep 02 15:37:47 bluelightning, I'm copying in a ubi file into the rootfs which I don't think is generic enough Sep 02 15:38:08 also all sorts of assumptions are made about bootloader file names (U-boot, uEnv, user.txt) etc Sep 02 15:38:44 hmm Sep 02 15:39:15 is this only likely to be applicable to one machine then? Sep 02 15:40:36 bluelightning, yeah. specially the bootloader environment files etc make assumptions about the machine (i2c bus names etc) Sep 02 15:41:43 bluelightning, also Richard was suggesting that this was attempted before but requires special mounts etc which is true Sep 02 15:42:02 we need it in there though so we could move it to the bsp for now Sep 02 15:42:06 sounds good? Sep 02 15:42:33 joelagnel: well as a temporary measure but really duplicating class files is worth avoiding Sep 02 15:42:45 just because of the pain of keeping it up-to-date Sep 02 15:43:01 sure Sep 02 15:43:06 I have to admit I'm unsure of the correct solution in this case though Sep 02 15:43:30 it's worth trying to coax people on the mailing list to come up with the right answer :) Sep 02 15:44:48 bluelightning, sure will do ,thanks :) Sep 02 15:47:40 03Simon Busch  07master * r7d8a59d98a 10openembedded.git/recipes/tasks/task-base.bb: Sep 02 15:47:40 task-base: bump PR to pull in changes to initscripts Sep 02 15:47:40 Signed-off-by: Simon Busch Sep 02 15:47:45 03Simon Busch  07master * r316e3b103d 10openembedded.git/recipes/freesmartphone/msmcomm.inc: Sep 02 15:47:45 msmcomm: bump SRCREV Sep 02 15:47:45 Signed-off-by: Simon Busch Sep 02 15:47:45 03Simon Busch  07master * r3ec6d684e8 10openembedded.git/recipes/initscripts/ (files/palmpre/umountfs initscripts_1.0.bb): Sep 02 15:47:45 initscripts: add a palmpre specific umountfs script Sep 02 15:47:45 Signed-off-by: Simon Busch Sep 02 16:14:32 48:5B:39:6F:53:77 Sep 02 16:14:36 ops sorry Sep 02 17:20:08 03Richard Purdie  07master * r2a7c92bdad 10bitbake.git/lib/bb/event.py: Sep 02 17:20:09 bitbake/event: Allow event handlers to quietly raise SkipPackage events Sep 02 17:20:09 Signed-off-by: Richard Purdie Sep 02 17:30:40 How should build-time checks for python packages be handled? Does the reciipe need to depend on python-numpy-native as well as python-numpy? **** ENDING LOGGING AT Sat Sep 03 02:59:56 2011